A potential therapeutic strategy for genetic diseases is to alter the genetic constitution of the affected tissues by means of grafts of normal precursor or stem cells. Over several years, evidence has accumulated to suggest that primary diseases of skeletal muscle, such as Duchenne muscular dystrophy, may be susceptible to this approach. This review makes a critical examination of such background evidence, and also of more recent data directly addressing the concept of therapy by means of grafts of normal myogenic cells. It is concluded that the data establish the principle that such grafts effect an alteration of the genetic constitution and phenotype of skeletal muscle and, therefore, might be used to alleviate recessively inherited myopathies. Several obstacles to the therapeutic application of this method to human disease are also identified; these seem to be problems of a technical nature rather than of basic principle, and none appears insuperable.